ArtificialWatch is an independent AI model release tracking service whose core function is to monitor the APIs of major AI providers and notify users immediately when new models go live.
Purpose Help developers, researchers, and AI enthusiasts learn about new AI model releases as soon as possible and avoid missing important updates.
Main capabilities
- Poll directly accessible provider APIs and public model catalogs every 60 seconds; the same model must be confirmed by two consecutive scans, and test prompts are used whenever possible to verify whether the model responds, in order to reduce false positives.
- After a release is confirmed, send alerts through the channels enabled by the user: free users receive Chrome desktop push notifications (within about one minute) and batched email (after about 15 minutes); paid users receive instant text messages (Signal, $12/month) or phone calls (The Call, $19.99/month).
- Provide a release radar showing confirmed and upcoming model releases, including leaks, rumors, and live Polymarket odds. The tracking list covers 172 models, such as Claude Opus, Grok 5, Gemini 5, DeepSeek V5, Kimi K3.1, and more.
- Watchtower ($34.99/month) provides prompt evaluation: up to 20 of the user's own prompts are run on the new model and the model currently in use, blindly judged by a third-party laboratory referee, with email conclusions and a PDF report.
- Subscription monitoring: track AI subscriptions the user already pays for (currently the Claude model lineup and the Gemini app model list, as well as each provider's API pricing and limits), and send alerts when models are added or removed, prices or limits change, or models are retired.
How to use Visit the website; after registering for free, you can choose the range of models to track (SOTA only, all major labs, or a custom list) and notification channels (Chrome push, email, text message, phone call). No dashboard maintenance is required; all monitoring runs automatically. Paid plans can be canceled at any time, and after cancellation you can still use them until the end of the paid period.
Use cases
- Developers who need to test new models as soon as possible.
- Researchers and technical decision-makers who follow AI model developments.
- Users who want to compare the performance of new models with the models they commonly use.
Limitations
- Alert confirmation depends on public API data; models that cannot be called for verification will be explained in the alert email.
- Phone and text message features are currently limited to paid plans.
- macOS menu bar, iOS/Android appsand other clients are still in planning and are not yet available.
- Free email is sent in batches, with a delay of about 15 minutes.
- Not affiliated with any AI laboratory.
- Specific platform support (such as Windows/Linux tray) and roadmap implementation timelines are not yet available.
